Rick Altherr wrote:
>
> On Apr 14, 2009, at 1:44 PM, Øyvind Harboe wrote:
>
>>> I'd rather know _why_ something failed rather than having to dig
>>> through the
>>> code to figure out which layer and why. Not every user is a UNIX
>>> programmer
>>> with intimate knowledge of the targets, interfaces, and general
>>> protocols.
>>
>> That's what the LOG_ERROR()'s are for. They tell you where and why.
>>
>> retval's are not propagated consistently or relyably in OpenOCD (or
>> any other C program I've seen really). We could switch to a language
>> with exceptions(C++), but we've had that discussion and there isn't
>> a strong incentive to do so with OpenOCD now that we've got exceptions
>> and resource tracking for menial stuff in Tcl.
>>
>>
>> --
>> Øyvind Harboe
>> PayBack incident management system
>> Reduce costs and increase quality, free Starter Edition
>> http://www.payback.no/index_en.html
>
>
> So because it isn't that way today, we shouldn't set a policy to do so
> in the future?
if( If Rick is speaking out for C++)
{
I second that.
Getting the existing code to compile with C++ is not a large
job. From that point, actually using C++'s useful features can be done
over time.
I dare say 2 days to get the existing code to compile with C++,
at most.
But I would not volunteer for this without write access to some
out of the way branch of the repo during the process.
}
C++ exceptions make error handling easier than C, plus you get type safe
linkage.
Dick
_______________________________________________
Openocd-development mailing list
Openocd-development@lists.berlios.de
https://lists.berlios.de/mailman/listinfo/openocd-development
